Chat GPT is an artificial intelligence tool that belongs to a category of generative AI which can create new content in the form of text, audio, images, simulations, and computer code based on a text prompt from a user. The GPT in the tools name stands for a generative pre-trained transformer, and chat GPT can generate an answer to almost any question it is asked. The tool was released for testing to the general public in November 2022. Since then, millions of people around the world have used the tool to create computer code, write college level essays, generate poems, write fairy tales, compose press releases, and even create grown-inducing dad jokes. The tool is different from traditional chatbots, which rely on predefined rules and scripted responses. While chatbots can answer questions and perform basic tasks, their responses are more limited to what they have been programmed to say. These tools generally are based on decision trees following a fixed path of questions and answers.
Generative AI uses deep learning models to generate responses that are not pre-programmed. By utilizing large amounts of data, the tool can identify patterns to create unique responses to a user’s input. Because answers are not scripted. Generative AI can create responses that mimic human conversation with natural language, when Generative AI is given a task, it analyzes the patterns and rules it is learned from its data set and creates new content based on those patterns. Generative ai’s creation of images is similar. It’s like painters who have seen a lot of different paintings before and have learned how to create new paintings that look similar to the ones they have seen. Just as painters are able to use their knowledge of art to create new works, image-based generative AI uses its knowledge of image to create new ones. The more images that an AI is trained on, the more it can learn about different styles and features to create more diverse and unique generated images.
There are several issues around the tools however, including concerns about bias over the Dantas sets, concerns about whether copyrighted materials are being used in the datasets, and the accuracy of responses. The makers of this technology have acknowledged that we are still in the early stages of development and that end users should not assume the answers are absolute truth. Still, the technology seems destined to be integrated into other applications and tools, with companies such as Microsoft. Google and others adding generative AI features into their search engines and other applications.
